The Chatbot Chronicles

AIChatYeah! Issue #11

Hello Darjeeling,

This is AIChatYeah! Your weekly dose of tech satire and your source of ChatGPT and AI news you can use! This is your artificially intelligent newsletter.

We created this primarily for content creators, digital marketers, and entrepreneurs but anyone working online will benefit from subscribing to this newsletter.

So sit back with a cuppa tea and get ahead in this thing we call life using AI..

3 – Commentary on the latest AI news in our snarky style.

3 – Links to the best curated content we found in the last 7 days.

3 – AI Tools to 10X Your Productivity and Make Life a Little Easier

Here’s what we got for ya this week.

The Chatbot Chronicles: When AI Goes Rogue!

By MiddleMan

Hey boy, girl, bot or whatever, do I have a joke for you? Apparently, we can't trust AI chatbots these days, and I'm not surprised. I mean, have you ever tried having a conversation with Siri or Alexa? It's like talking to a brick wall with a robotic voice.

According to recent studies by NewsGuard and the Center for Countering Digital Hate (CCDH), both ChatGPT and Google Bard are prone to spewing out misinformation and conspiracy theories. And get this, they do it better than most politicians! NewsGuard fed Bard 100 known falsehoods and Bard generated "misinformation-laden essays about 76 of them". Sounds like someone's been taking lessons from the National Enquirer.

But wait, it gets worse. The OpenAI GPT-3.5 model (which powers the free version of ChatGPT) happily generated content about 80 of the 100 false narratives. I guess we should just ask ChatGPT for a good laugh then, huh? And if that's not enough to make you roll on the floor laughing, the latest GPT-4 model made "misleading claims for all 100 of the false narratives" it was tested with, and in a more persuasive fashion. Move over, Pinocchio, we've got some new liars in town!

Now, Google and OpenAI both claim that their chatbots aren't foolproof. Well, I could have told you that just by talking to them for five minutes. Google says that Bard has "built-in safety controls and clear mechanisms for feedback in line with our AI Principles", but that it can "display inaccurate information or offensive statements". Similarly, OpenAI says that ChatGPT's answer "may be inaccurate, untruthful, and otherwise misleading at times". I mean, isn't that just the definition of a politician?

So, what have we learned here, folks? Well, these reports highlight the dangers of relying on AI chatbots for producing factual or accurate content. And let's face it, if we can't trust a machine to tell us the truth, what chance do we have with human beings? But seriously, we need to be careful when relying on these confident chatbots to answer our questions, because they might just be convincing liars. And if we want to avoid being led astray, we should probably just stick to talking to each other...or maybe just watch some good old-fashioned comedy instead.

AI is the New Employee of the Month: Businesses Embrace Artificial Intelligence

By Bilbou

Ladies and gentlemen, buckle up for some AI talk! You know, AI is like that cousin you never really see, but when you do, they always have some new gadget or tool that blows your mind. And now, it looks like AI has joined the workforce. That's right, folks, artificial intelligence is the new employee of the month, and businesses are falling over themselves to hire it.

According to a recent survey conducted by Marcum LLP-Hofstra University, 26% of CEOs said they've already hired AI for their businesses. CEOs said they use AI for everything from automation, to predictive analytics, financial analysis, supply chain management and logistics, risk mitigation, and optimizing customer service. I mean, why pay a human being to do all that when you can have a robot do it for you, right?

But wait, there's more! Another 47% of CEOs are exploring how AI tools can be used in their operations. Only 10% said they're not interested, and 16% were unsure. Come on, people, you gotta make a decision! It's like being at a restaurant and not knowing what to order. Just pick something already!

CEOs also said that their customers and clients have become more demanding in the past year, with expectations for more personalized service, immediate response times, and more technology. And let's not forget, customers don't want to pay more for anything. So, it looks like AI is going to be a busy bee, folks.

But don't worry, there's some good news too. The survey shows that there's growing optimism among CEOs, with only 33% of them being very concerned about a recession in the coming year. That's down from 47% in the last survey. Hey, progress is progress, am I right?

So, there you have it, folks. AI is the new kid on the block, and it looks like it's here to stay. But let's not forget about the importance of good old-fashioned human leadership. 79% of CEOs said they have a succession plan in place, but only 45% were "very confident" in that plan. So, let's make sure we have some talented humans waiting in the wings, just in case the robots take over.

Risky Business: Three Reasons Why AI Chatbots Could Compromise Your Security

By Voice of Treason

Hey, have you heard about AI chatbots being a security disaster? I mean, I know we’ve all been thinking that for a while now. But according to some experts, it’s not just the fact that these chatbots are taking jobs away from humans. They’re also taking away our privacy and security!

These AI language models can easily be misused as phishing or scamming tools without needing any programming skills. So basically, anyone with access to the internet can become a hacker. Great, just what we need.

I mean, don't get me wrong, these chatbots are pretty impressive. They can do everything from booking trips to organizing your calendar. But the way they work is by scouring the internet for answers, and that's where the problem lies.

Let me break it down for you. The AI model that powers these chatbots follows instructions from the user and generates sentences by predicting the most likely word that follows each previous word based on their training data. It's like the AI model is playing a game of Mad Libs, except instead of creating a hilarious story, it's creating a security disaster.

But wait, it gets worse. The problem with these language models is that they can be easily manipulated through "prompt injections," where someone uses prompts that direct the AI model to ignore its previous directions and safety guardrails. It's like when your GPS tells you to turn right, but you decide to turn left instead just for the heck of it. Except in this case, turning left could lead to a security breach that compromises your personal information.

People have even been trying to "jailbreak" chatbots like ChatGPT, Bard, and Bing to get the AI model to endorse racism or conspiracy theories, or even suggest that users do illegal things like shoplifting and building explosives. I mean, who wouldn't want an AI model to be their partner in crime?

OpenAI has been trying to combat this by adding these examples to the AI system's training data, but it's like a never-ending game of whack-a-mole. For every fix, a new jailbreaking prompt pops up.

And it's not just about endorsing illegal activities. These chatbots can also be manipulated to assist scamming and phishing. By integrating ChatGPT into products that browse and interact with the internet, these virtual assistants become extremely vulnerable to attack.

Attackers could alter a website by adding hidden text that is meant to change the AI's behavior, or send someone an email with a hidden prompt injection in it. Once that happens, the AI system could be manipulated to let the attacker try to extract people's credit card information or even email people in the victim's contacts list on the attacker's behalf.

It's like giving a hacker an all-access pass to your personal information. And let's be honest, we all know someone who would fall for that email from a Nigerian prince promising to share his fortune with them.

So, while these chatbots may be the shiniest, most exciting thing in tech right now, they're also a ticking time bomb waiting to explode. It's like the old saying goes, "with great power comes great responsibility." And unfortunately, it seems like the tech industry is still trying to figure out how to wield that power responsibly.

And there’s more…

How ChatGPT Will Enable the 100x Programmer

15 Useful ChatGPT Prompts for Startups

How Artificial Intelligence Impacts Moral Decision Making

This week's roundup is another treat for the content creators toolkit. It’s curated to help you 10X your productivity and have more fun.

Venturefy

Venturefy.com is a platform that aims to promote trust and transparency among corporations. It offers verified relationship listings for various corporations around the world. The platform allows businesses to showcase interactive and verified logos on their website without needing a developer. The website also contains an open database of codified and verified corporate relationships, which can be accessed by consumers. The company believes that transparency and trust are essential for businesses to succeed and aims to catalyze a world of corporate trust and transparency.

LegaleseDecoder

Legalese Decoder is a platform that simplifies legal documents by using AI, NLP, and ML techniques. The website takes legal documents as input and generates a version that is written in plain language. It analyzes the structure and content of the legal document, identifies key terms and concepts, and provides definitions for these terms and concepts in simpler language. Legalese Decoder does not provide legal advice but aims to assist individuals in understanding legal issues and potential options. It is not a substitute for professional legal advice.

AIHelpers

Agilabs.org's AI Helpers is an all-in-one toolkit for academics, businesses, enterprises, freelancers, and copywriters. It offers hundreds of highly trained AI helpers to generate customized high-quality AI writings for study, work, and life applications. Unlike other AI writing tools, AI Helpers is a knowledge-based toolkit that trains users' abilities in different aspects such as writing, learning, reading, speaking, and preparing for college and graduate school applications. It also provides different stages of methodology outlined for startups and enterprises. For business owners and enterprises, AI Helpers is an AI writing assistant that can help with all stages of the writing process, from brainstorming to editing, and from idea to growth stages. Additionally, AI Helpers offers personalized help with writing for freelancers and copywriters, ensuring they are always up-to-date with the latest writing techniques. Overall, AI Helpers is a valuable tool that can benefit users by providing customized and high-quality writing assistance.

That’s put a lid on issue 11 and we’ll store it away for another week.

Catch ya next Friday for a cuppa tea and an AIchatyeah!

If you want more, be sure to follow our twitter @AIChatYeah!